The internet has become a universal platform for web communications. Technologies for optimizing web delivery of content have been developed. Various delivery methods are available as they rely on standard web servers that are widely supported. For instance, a content server for a content delivery network can break a media stream into small fragments and transfer the fragments to client devices. The client devices download the files and play the fragments in an order specified by the content server. When too many client devices request data from the content server at the same time, the existing bandwidth and system resources of the content server can be overwhelmed.
A content delivery network may include edge servers that are connected to the content server to mitigate the problem. The edge servers can aggregate requests by multiplexing connections from the client devices to one connection between each edge server and the content server. The edge servers can receive corresponding requested data from the server. In response to download requests from the client devices, the edge servers relay the corresponding requested data to the client devices. The edge servers can further cache the data such that cached data can be made available to further data requests from client devices connected to the edge servers.
However, just like the connection limitation between the content server and a client device, there may be a connection limitation between the edge server and the client device as well. Various factors can limit the bandwidth for the client device to download data from the edge server. For instance, the client device may connect to the Internet via a network provider with a data transferring speed upper limit. Therefore the maximum download speed from the edge server to the client device is limited by the speed upper limit.